Advances in statistics
This article is part of a larger document. View the larger document here.Abstract
The papers included in the Advances in Statistics section of the Partners in Flight (PIF) 2002 Proceedings represent a small sample of statistical topics of current importance to Partners In Flight research scientists: hierarchical modeling, estimation of detection probabilities, and Bayesian applications. Sauer et al. (this volume) examines a hierarchical model describing attributes and change of aggregates of bird populations using a Bayesian statistical inference approach with WinBUGS. They point out that Win- BUGS provides a user-friendly software environment for hierarchical modeling that more realistically describes many biological contexts. The authors of two papers focus on the issue of detection. Farnsworth et al. (this volume) presents a summary of detection probability estimation procedures, including distance sampling, double observer methods, time-depletion (removal) methods, and hybrid methods that combine these approaches. They conclude by presenting a method that combines distance and removal sampling methods, along with results. Earnst and Heltzel (this volume) describe estimates of detection ratios based upon songbird surveys as a function of species, forest type, and season. They report that detection ratios reflecting detectability differed with species (not surprising), but also with timing of surveys (even a couple of weeks makes a difference) and habitat. Stauffer et al. (this volume) describes a Bayesian interpretation of Akaike weights, useful for assessing the competitiveness of a collection of models with a series of datasets. They illustrate these ideas with habitat selection models for Northern Spotted Owl in California.